The object of the proposed research is to test the feasibility of using memory for prose in diagnosis and in the understanding of thought processes in clinical populations. Where a marked onset of symptoms occurs, prose assummed to have been learned before that onset, as well as prose learned after that onset will be contrasted. Quantitative descriptions of which parts of standard test passages different clinical populations remembered will be compared.